Base64 Image Compressor

Base64 is encoding, not image compression

Standard Base64 turns each three input bytes into four text characters, so its length is 4 × ceil(n / 3) and the long-run overhead is about 33.3%. A data URL adds a MIME prefix on top of those characters.

This tool reduces the image bytes first: it decodes the image, changes dimensions, quality, or format when useful, and then creates standard Base64 again. The result needs no custom decompressor.

Why Gzip is a different output

Gzipping Base64 produces compressed binary data that must be decompressed before an image element can read it. Encoding that Gzip stream again does not make it a normal image data URL.

JPEG, PNG, and WebP are already compressed formats. Canvas re-encoding may help by reducing pixels, changing a lossy quality, removing metadata, or choosing another format, but it is not guaranteed to be lossless or smaller.

Target size without returning a larger string

For JPEG and WebP, the tool searches quality with a fixed iteration limit. If the lowest useful quality is still too large, it reduces dimensions in bounded steps while preserving aspect ratio.

If no valid result has fewer standard Base64 characters, the original bytes are kept. A near-target result is clearly marked when the requested KB limit cannot be reached.

Local processing and format boundaries

The image stays in this browser tab. No file upload, account, remote URL, telemetry, or third-party image library is used. The 20 MiB decoded-byte and 16 megapixel limits protect the tab from accidental memory pressure.

SVG, GIF, ICO, and animated WebP are rejected rather than executed or silently flattened. Transparency is preserved in Auto mode; converting transparent pixels to JPEG requires an explicit background color. Canvas re-encoding is not a malware scan.

When not to use Base64

Data URLs are useful for small self-contained assets, fixtures, and paste-ready API fields. Large images usually work better as Blob URLs, separate files, or binary uploads because they can be cached and streamed independently.

Use the byte and Base64 comparisons below to judge the real result. Removing a data: prefix is never counted as compression.

Common failures this tool explains

• Malformed Base64, impossible padding, or characters outside the Base64 alphabets.
• A declared data URL MIME type does not match the decoded magic bytes.
• The image exceeds 20 MiB, 16 megapixels, or the browser cannot decode it.
• The source is animated, multi-frame, SVG, or already too optimized to become smaller.

Equivalent commands

Use the runtime or shell already available in your workflow. Validate untrusted input and keep binary output out of terminals.

from PIL import Image
from io import BytesIO
import base64
image = Image.open(BytesIO(base64.b64decode(value)))
out = BytesIO(); image.save(out, 'WEBP', quality=82)
result = base64.b64encode(out.getvalue()).decode('ascii')

const blob = new Blob([bytes], { type: mime });
const bitmap = await createImageBitmap(blob);
ctx.drawImage(bitmap, 0, 0, width, height);
const output = await new Promise(r => canvas.toBlob(r, 'image/webp', 0.82));
bitmap.close();

printf '%s' "$VALUE" | base64 --decode > input.png
magick input.png -resize '1920x1920>' -quality 82 output.webp
base64 --wrap=0 output.webp